Facilitators Associated With Building and Sustaining Therapeutic Alliance in Advanced Pediatric Cancer

IMPORTANCE Therapeutic alliance is a core component of patient- and family-centered care, particularly in the setting of advancing cancer. Communication approaches used by pediatric oncologists to foster therapeutic alliance with children with cancer and their families are not well understood. OBJECTIVES To identify key oncologist-driven facilitators associated with building and sustaining therapeutic alliance in the setting of advancing pediatric cancer and to develop a framework to guide clinical practice and future investigation of therapeutic alliance. DESIGN, SETTING, AND PARTICIPANTS In this qualitative study, serial disease reevaluation discussions that occurred in the clinic, inpatient hospital, or off campus via telephone were recorded among pediatric oncologists, patients with high-risk cancer, and their families across 24 months or until death, whichever occurred first, from 2016 to 2020. This analysis focused on recorded discussions for pediatric patients who experienced progressive disease during the study period. Content analysis was conducted across recorded dialogue to derive inductive codes and identify themes. Participants were patient-parent dyads for whom a primary oncologist projected the patient’s survival to be 50% or less, all family members and friends who attended any of their recorded disease reevaluation conversations, and their oncologists and other clinicians who attended the recorded discussions. RESULTS A total of 33 patient-parent dyads were enrolled and followed longitudinally. From this cohort, 17 patients experienced disease progression during the study period, most of whom were female (11 [64.7%]) and White (15 [88.2%]) individuals. For these patients, 141 disease reevaluation discussions were audio recorded, comprising 2400 minutes of medical dialogue. Most children (14 [82.4%]) died during the study period. A median of 7 disease reevaluation discussions per patient (range, 1-19) were recorded. Content analysis yielded 28 unique concepts associated with therapeutic alliance fostered by oncologist characterize the to to in with intention. Recognizing that personalities and communication styles naturally vary, clinicians may review the framework and choose to highlight behaviors that feel most genuine to them. We studied conversations recorded at times of crisis for patients and families, and even in these painful spaces, we found moments wherein clinicians created meaning (even laughter) through connection and caring. We hope that the facilitators underpinning this framework may help guide clinicians in forming meaningful therapeutic bonds, particularly when sharing difficult information, as a way to affirm our shared humanity and communicate to patients and families that their feelings matter. This study also highlights the need for future investigation of several important, unexplored questions. We observed interactions among oncologists, patients, and their families across evolving illness, and we identified potential facilitators associated with therapeutic alliance. However, the identified concepts reflected our perspectives as clinicians and observers. We did not ask parents and children what they personally valued with respect to therapeutic alliance, and we do not know whether these interactions actually supported the human bond between clinicians and children or their families. We also do not know how cultural background or personal lived experiences may influence oncologists’ comfort levels with using certain facilitators of relationship building. For example, although the concept of “sharing personal information” emerged as a frequent approach, this method may not be appropriate or valued by everyone. Similarly, the act of “giving opinions” may come across as paternalistic by some stakeholders. The constructs that we identified, informing development of this initial theoretical framework, require validation to determine whether they align with what patients and families believe to be most important. Future investigation should consider participatory research methods to engage patients and families in assessing the trustworthiness of constructs and prioritizing communication approaches that they perceive to be most meaningful. the illness and care experience for children with cancer andtheirfamilies.Drawingfromlongitudinalmedicaldialogueacrossanadvancingillnesscourse,we identified 7 core themes associated with therapeutic alliance: human connection, empathy, presence, partnering, inclusivity, humor, and honesty. These findings offer a framework to support clinician education and future opportunities to study key stakeholder perspectives on approaches to optimize therapeutic alliance among oncologists, patients, and their families.


Introduction
Patients with cancer and their families experience profound physical, psychological, and existential suffering across the illness trajectory. [1][2][3][4] Development of therapeutic alliance between the patient or family and the oncologist may help ease the pain. [5][6][7] Therapeutic alliance refers to the nature and strength of an affective bond between the patient or family and the clinician in collaboration toward shared goals. [8][9][10][11] Among adult patients with advanced cancer, a strong therapeutic alliance with the oncologist has been associated with emotional acceptance of terminal illness and decreased invasive interventions at end of life. 5 For young adults with cancer, the patient-oncologist alliance also has been shown to be protective against suicidal ideation. 12 In addition, a positive patient-oncologist therapeutic alliance appears to benefit caregivers of adult patients with cancer across the illness course and during bereavement; more specifically, a patient's report of a strong therapeutic alliance has been shown to be associated with a caregiver's report of better social function, health-related quality of life, and emotional well-being following the patient's death. 6 As proposed by the National Cancer Institute 13 and further validated in pediatric cancer, 14 relationship building is considered one of the core functions of patient-centered communication. 15 Most parents of children with cancer form trusting relationships with their child's oncologist, and development of trust is associated with parent perception of high-quality physician communication. 16 Fostering therapeutic alliance through patient-centered communication is particularly critical in the context of cancer relapse or progression, when children and families navigate stressful decisions. 17 Unfortunately, poor outcomes have been shown to threaten therapeutic alliance, with parents reporting decreased trust in oncologists in the setting of advancing disease. 16 Although high-quality oncologist communication is known to be important for therapeutic alliance, 16 specific communication approaches used by oncologists to facilitate therapeutic alliance in pediatric cancer are not well understood. The U-CHAT (Understanding Communication in Healthcare to Achieve Trust) trial was designed to better understand the evolution and impact of communication across the advancing pediatric cancer trajectory. 18 In this prospective, longitudinal study, serial disease reevaluation conversations among pediatric oncologists, patients with high-risk cancer, and their families were audio recorded across the illness course to the time of death. The primary aim of this study was to characterize the evolution of communication strategies used by pediatric oncologists across advancing illness. In this article, we present the findings from a secondary aim of that study: to identify and describe oncologist approaches for promoting therapeutic alliance, with the goal of developing a framework to organize emerging constructs and better understand potential facilitators for alliance to target in future research.

Methods
The protocol for this qualitative study was developed by an interdisciplinary team of pediatric oncology and hospice and palliative medicine experts in collaboration with a bereaved parent steering council; it was reviewed and approved by the institutional review board at St Jude Children's Research Hospital in Memphis, Tennessee. Data were collected between 2016 and 2020. We present study methods and findings following the Consolidated Criteria for Reporting Qualitative Research (COREQ) reporting guideline and checklist. 19 All participants provided written or verbal informed consent that was obtained in a manner consistent with the Common Rule requirements. No one received compensation or was offered any incentive for participating in this study.
Details about the study protocol and feasibility and acceptability metrics for enrollment and capture of longitudinal data have been previously published. 18,20 In brief, we enrolled a convenience sample of 6 pediatric oncologists who provided care to patients with solid tumors at an academic pediatric cancer center and obtained verbal informed consent. We then enrolled 4 to 6 patientparent dyads with poor prognoses per oncologist. Eligible patients were aged 0 to 30 years, diagnosed as having non-central nervous system solid tumors, and considered to have poor prognosis, with the primary oncologist estimating survival of 50% or lower. Eligible parents or guardians were legal caregivers aged 18 years or older with English language proficiency who planned to be present for disease reevaluation discussions. Eligible patient-parent dyads were identified by the research team through review of outpatient clinic schedules and institutional trial lists, and permission to approach dyads was requested from the primary oncologist. A research team member obtained written informed consent from patients and caregivers during a clinic visit that did not include a disease reevaluation discussion; patients aged 12 years or older provided formal assent, and patients aged 18 years or older and parents provided consent. Demographic and disease-related information were extracted from the electronic medical record.
Patient-parent dyads were followed prospectively, with serial disease reevaluation discussions among patients, their family members and friends, and oncologists and other clinicians recorded in the clinic, hospital, or off campus via telephone across the illness course until death or 24 months from disease progression on study, whichever occurred first. The present work focuses on analysis of recorded discussions for dyads who experienced progressive disease during the study period.
To characterize key components of therapeutic alliance building in advancing pediatric cancer, we performed content analysis of recorded medical dialogue. We wanted to approach the data with openness to allow the dialogue to inform the findings; however, we also started with a priori constraints that shaped the lens through which we assessed the data (ie, we wanted to focus on clinician behaviors that served as potential facilitators of therapeutic alliance), precluding use of a traditional grounded theory approach. A research team representing medical and nursing perspectives across pediatric oncology and palliative medicine ( Table 1) reviewed the literature related to therapeutic alliance in oncology; finding little consensus for fundamentals of alliance building in pediatric cancer care, we used an inductive approach 21 to identify oncologist communication approaches to foster affective bonds with patients and families. Two researchers (E.C.K. and S.R.) repetitively listened to audio recordings, conducted extensive memo writing, and used raw data to inform development of codes, code definitions, and salient examples related to therapeutic alliance building. 22 Additional researchers (C.W. and J.N.B.) reviewed recorded content and provided feedback in iterative cycles of codebook development. As facilitators were identified, we reviewed existing frameworks when possible to help codify concepts. For example, for expressions of empathy, we referenced the naming, understanding, respecting, supporting, and exploring framework for navigating emotions. 23,24 However, through the processes of memo writing and inductive code development, we created definitions for each code based on the concepts emerging from the dialogue. The codebook was finalized following deep review of sufficient raw data to reach saturation, with no new therapeutic alliance concepts emerging from the recorded dialogue.
The codebook is presented in Table 2.

Data Analysis
Coding processes were conducted within MAXQDA, a mixed-methods data analysis software system. 25 Three analysts (E.C.K., S.R., and C.W.) pilot tested the codebook across a series of medical dialogue recordings to identify areas of variance. All analysts met to reconcile variances and achieve consensus, modifying the codebook as needed to improve dependability, confirmability, and credibility of independent codes. 26 Following codebook finalization, independent coding was performed by 2 analysts (S.R. and C.W.), with weekly meetings to review any coding variances and third-party (E.C.K. and J.N.B.) adjudication to reach consensus. Consistency in code segmentation also was reviewed to ensure a standardized approach (E.C.K., S.R., and C.W.). Code frequency, temporal duration, and distribution across discussion subtype were calculated and reported as

Results
In total, 33 patient-parent dyads were enrolled and followed longitudinally. Data on patient-parent dyads who declined enrollment in the larger study have been published previously. 20 In brief, 7 of 41 approached dyads (17%) did not enroll owing to hesitation or refusal by either the patient (n = 4) or parent (n = 4). Although these numbers are small, refusal rates did not appear to disproportionately exclude dyads based on race and ethnicity. Only 1 of the 7 dyads selfidentified as Black (approximately 14%) and 1 as Hispanic (approximately 14%), which are roughly equivalent to the percentages of Black patients and Hispanic patients treated at the institution.
Across the advancing illness course, we identified 28 unique concepts as potential facilitators associated with therapeutic alliance, with further synthesis generating 7 core themes: human connection, empathy, presence, partnering, inclusivity, humor, and honesty ( Table 4). We present these themes as a framework for potential facilitators associated with therapeutic alliance in pediatric cancer (Figure).

Human Connection
Oncologists promoted connection on a human level with the patient and family through 4 approaches. In the remembering approach, oncologists recalled information, unprompted, related to the patient's or family's personal life. Some recollections were lighthearted or fun ("Where is my birthday twin?"); other times, oncologists recalled prior feelings central to the patient's values related to decision-making ("I know that in the past, you've also said it's very hard to be away from family.").
Through sharing, oncologists also contributed personal information about themselves or their lives as an approach to find common ground with the patient or family. Shared information ranged from minor details about their personality or likes or dislikes (eg, "I'm not a reptile person.") to a window  Oncologist links patient's symptoms or pain with scan results or disease progression to provide clarity or understandable medical information Using analogy Oncologist uses an analogy or a prop to provide clarity or understandable medical information Showing images Oncologist shows the patient or family the imaging findings (or will show in the near future) to provide clarity or make medical information more understandable

Comedy
Oncologist use of comedic relief, attempt at humor, and joking during conversations

Ribbing
The use of playful teasing by the oncologist

Matching maturity level
Oncologist matches the tone, language, and maturity level of a patient to connect with them

Honesty
Warning shot Oncologist opens with a statement that gives patient or family a moment to emotionally prepare for hearing bad news Transparency Oncologist uses statements that attempt to transmit or highlight realistic prognostic assessment related to delivering good or bad prognosis; a linguistic choice that captures oncologist attempt to bond through transparency; includes language about being honest or "I worry" statements Giving opinion Oncologist uses statements of ownership, including "I think," "I feel," "I recommend," or "I believe," or synonyms, while discussing illness course, treatment plan, goals of care; statements that show ownership or personalize opinion while building alliance or partnership with patient or family; includes any time the oncologist uses the phrase "If it were me or my child…"

Summarizing
The oncologist uses a summary statement to reiterate results, treatment, or findings of scans into their personal life or emotions (eg, "I don't know about you guys, but the uncertainty drives me crazy."). Oncologists used friendly conversation or small talk unrelated to the illness or treatment plan to connect with patients and families. Finally, affection or expressions of fondness were also used to help generate connection on a human level: "It's okay if you don't want to, sweetheart."

Empathy
We identified 7 facilitators of empathy, including 2 de novo concepts and 5 concepts originating from naming, understanding, respecting, supporting, and exploring, each of which was refined during the inductive coding process. In the first de novo concept, standing in another's shoes, oncologists used empathetic statements to convey a willingness to imagine the patient's or family's struggle: "You guys are so wonderful. And I can only imagine how hard this is too because you've lost a lot of friends." In the second de novo concept, saying sorry, oncologists used "I'm sorry" statements as a  We're not allowing, we're not allowing her to leave us unless, you know, unless she's getting something and promises to come back. • But there's no way to ever know until we go for it, right? We go as hard as we can and nobody's stopping that. So I don't want you to feel at all like anybody is saying, "Oh, we're done and we're not going to keep trying." No. We're going to keep trying for sure. Without fail, okay? Team mentality • I want you to live, and I want you to graduate. And that's why we will do whatever you want us to do as much as you want us to do to try to make that happen. • It would be nice if we got the next set of scans and things were the same or slightly better. You know, we'll take whatever we can get. • What I would be happy to do today is to say that we probably shouldn't make a decision today.
That I think we should take a little time.
(continued)  • The biggest one is this one, which is why she complains here. But this one, this one, this one, this one, and this one, and you said that sometimes she complains of pain here, that's exactly where it is. • I think it likely is from one of the, from that area over here, cause I think, just based on the location that you're reporting, it looks fairly consistent with that region. • There's also some uptake that's up in the skull base. Umm, and you know, it's possible that could be contributing to some of the headaches that you're having because you're saying that they're starting at the back of your head. , the bottom line from the scans that we just got is that we're seeing evidence of progression and that we're going to come off the study, okay? • We'll do everything all over again and then, you know, periodically for a couple years to make sure. because it still could come back.… As best we know, if 1 cell escapes all this, it eventually comes back, and that's the scary part. • I worry that, because of all these people here, you're worried more about them than [about] you. Are you? "That's a lot of information."] "It is a lot of information." Oncologists showed respect through provision of encouragement, praise, or reassurance: "You did everything right." Respect often overlapped with the concept of supporting, in which oncologists expressed concern alongside a willingness to help: "As long as you want to try something, we're willing to try something."

JAMA Network Open | Pediatrics
Oncologists also cultivated empathy through the concept of exploring, in which they offered patients or family members an opportunity to talk about whatever they were feeling or processing: "What do you want to, what's running through your head?"

Presence
We found 2 concepts that promoted a sense of meaningful presence as part of nurturing therapeutic alliance: being in the moment and silence. Oncologists verbalized with intention their willingness to stay in the moment, whether by giving time (eg, "We need to take the time. I think you guys deserve the time") or by silencing other distractions (eg, "I turned my ringer off"). Silence was also used with intention to create space for processing and emotional connection.

Partnering
We discerned 3 concepts that fostered a sense of partnership among oncologists, patients, and their families. In the concept of nonabandonment, oncologists expressed a commitment to stay, no matter how difficult things became: "You know that you will never shake us, like good luck trying." This concept of being "in it for the long run" also manifested through running imagery: "We love you, and we're going, we're all going to do this together, okay. This is a marathon and we're all going to just keep running." In the concept of team mentality, oncologists used "we" language to demonstrate a sense of teamwork and collaboration with respect to values, goals, or decision-making: "This is not a conversation that says that we don't keep fighting because we always keep fighting." Finally, in the concept of accommodation, oncologists partnered with patients and families by anticipating and meeting their needs and preferences. For example, the oncologists offered to reschedule visits or therapies to optimize quality of life: "I think we can do [the treatment] in a way that's not going to interfere with the trips that you have planned."

Inclusivity
We found 5 concepts that centered or included the patient or family with respect to transfer of knowledge or navigation of decision-making. In the open door concept, oncologists used open-ended • So that's a long way of saying, yes, I'll consider it, you know, with the caveat of that we gotta get to a state where we feel it's even worth your time. • And so to me that means that the tumor cells are dying, right? They're not as active, they're not replicating as much and growing, and so that's a good thing.
statements to lay groundwork for sensitive conversations and to prompt patients or families to voice their hopes, wishes, opinions, or goals: "If that's part of our goal . . . maybe we want to do that without getting into really intense things that are going to get you stuck in the hospital and things like that. You know, to me that seems like it would be a very, that would be a very reasonable goal, a good goal to have." Oncologists also reinforced inclusivity by affirming the voice of the patient or parent as integral to decision-making: "Buddy, every decision that we make is the right decision for you because they're your decisions, you know. We're not, you know, whatever you decide is not ever going to be the wrong choice." Practically, oncologists used several concrete strategies to center the patient or family in medical dialogue. By connecting symptoms, they linked imaging findings with how a patient felt, translating abstract data into meaningful patient-centered information: "The biggest issues are in your bones. I think that's why you feel bad." By using analogies, oncologists simplified medical information into lay language for patients and families: "Your bone can look a little bit more like Swiss cheese, to use an analogy, and then the trouble with Swiss cheese is it has holes, and so if pushing on it having a lot of stress you can sometimes get a little bit of collapse." Oncologists further included patients and families by showing imaging during discussions: "All this area over here is in her right shoulder, even though we radiated it, it's much, much worse. All this super black stuff, that is disease."

Humor
We identified 3 concepts related to humor as a key aspect of therapeutic alliance. In the concept of Oncologists develop and deepen therapeutic alliance through 7 thematic approaches: human connection, empathy, presence, partnering, inclusivity, humor, and honesty. We present a framework for conceptualizing the interconnectivity and interdependency of these core facilitators associated with therapeutic alliance in pediatric cancer. The framework aesthetic intentionally resembles a wheel, evoking the idea that therapeutic alliance helps to move communication forward more productively, enabling clinicians, patients, and families to travel collaboratively across the illness course. by aligning their tone or language choices with the patient's developmental stage: "I know that's what you're thinking, isn't it? Yeah, like, loser, why haven't we talked about this before?"

Honesty
We identified 4 concepts around honesty as a strategy to engender trust. In the warning shot concept, oncologists prefaced bad news with a forewarning to allow a patient or family to emotionally prepare: "So I'm afraid I don't have very good news for you. So let me start from the beginning, okay." In the concept of transparency, oncologists emphasized a desire to be honest or frank, particularly when discussing difficult information: "But it doesn't mean that we don't have plans. Okay, I just want to be honest." Phrases such as "I worry" also enabled oncologists to share their thoughts candidly. Oncologists further fostered alliance via the giving opinions concept, using statements of ownership (eg, "I think," "I feel," "I recommend," and "I believe") when discussing illness progression, treatment planning, or goals of care: "I think we need to continue this therapy. Whether it's really doing a lot or not, I don't know, but I feel very uncomfortable stopping." In the concept of summarizing, oncologists synthesized challenging information, thereby strengthening its trustworthiness: "What I would say, then, to kind of put the whole picture together, is that things are stable."

Discussion
Therapeutic alliance is one core component of patient-and family-centered communication, particularly in the context of advancing illness. Through analysis of key conversations among pediatric oncologists, patients with advancing cancer, and their families, we identified 7 themes informing a framework for potential facilitators associated with therapeutic alliance: human connection, empathy, presence, partnering, inclusivity, humor, and honesty (Figure). This framework has possible utility in clinical practice and future research. When striving to create, sustain, or repair therapeutic alliance, particularly during stressful time points across an advancing illness course, clinicians might draw from this framework to emphasize certain facilitators with intention.
Recognizing that personalities and communication styles naturally vary, clinicians may review the framework and choose to highlight behaviors that feel most genuine to them. We studied conversations recorded at times of crisis for patients and families, and even in these painful spaces, we found moments wherein clinicians created meaning (even laughter) through connection and caring. We hope that the facilitators underpinning this framework may help guide clinicians in forming meaningful therapeutic bonds, particularly when sharing difficult information, as a way to affirm our shared humanity and communicate to patients and families that their feelings matter.
This study also highlights the need for future investigation of several important, unexplored questions. We observed interactions among oncologists, patients, and their families across evolving illness, and we identified potential facilitators associated with therapeutic alliance. However, the identified concepts reflected our perspectives as clinicians and observers. We did not ask parents and children what they personally valued with respect to therapeutic alliance, and we do not know whether these interactions actually supported the human bond between clinicians and children or their families. We also do not know how cultural background or personal lived experiences may influence oncologists' comfort levels with using certain facilitators of relationship building. For example, although the concept of "sharing personal information" emerged as a frequent approach, this method may not be appropriate or valued by everyone. Similarly, the act of "giving opinions" may come across as paternalistic by some stakeholders. The constructs that we identified, informing development of this initial theoretical framework, require validation to determine whether they align with what patients and families believe to be most important. Future investigation should consider participatory research methods to engage patients and families in assessing the trustworthiness of constructs and prioritizing communication approaches that they perceive to be most meaningful.

JAMA Network Open | Pediatrics
Building and Sustaining Therapeutic Alliance in Advanced Pediatric Cancer Although preliminary, this framework outlines key components associated with creating therapeutic alliance, filling a gap in the literature. Few existing conceptual frameworks characterize the variables intrinsic to development and evolution of therapeutic alliance in pediatric cancer care.
One model derived from the experience of oncologists proposes a cyclical reinforcing association between effective communication and therapeutic alliance across the illness trajectory 15 ; however, that model does not propose constructs for building alliance. Several models for therapeutic alliance in psychology and psychotherapy exist, yet their authors acknowledge ongoing challenges with application of the models toward measurement and characterization of therapeutic alliance in medicine. 27,28 Several quantitative tools have been validated to measure therapeutic alliance between patients and oncologists. For example, The Human Connection scale is a psychometrically validated tool with Likert scale items that measures the extent of patient-reported mutual understanding, caring, and trust with their physicians, which has shown reliability in adult cancer populations. 5 Yet given the nuance and complexity of therapeutic alliance, a mixed-methods approach that incorporates qualitative analysis has potential utility. Recognizing the potential role that therapeutic alliance plays in influencing patient care, caregiver well-being, and bereavement outcomes, studying therapeutic alliance as a potential modifiable factor is important. This theoretical framework offers potential infrastructure to further our understanding of therapeutic alliance in future cancer research.

Limitations
Study limitations include single-site design and potential sample bias for patient-parent dyads inclined to value or promote therapeutic alliance; all eligible oncologists participated, mitigating selection bias for physicians. Racial and ethnic diversity was limited and requires prioritization in future work. Rarely, discussions were not recorded due to logistical issues or at the request of the participating patient or parent; missing data might influence concept generation. However, given the saturation of themes across thousands of recorded minutes, several missing time points are less likely to influence the synthesis of findings.

Conclusions
Therapeutic alliance is an important aspect of the illness and care experience for children with cancer and their families. Drawing from longitudinal medical dialogue across an advancing illness course, we identified 7 core themes associated with therapeutic alliance: human connection, empathy, presence, partnering, inclusivity, humor, and honesty. These findings offer a framework to support clinician education and future opportunities to study key stakeholder perspectives on approaches to optimize therapeutic alliance among oncologists, patients, and their families.